My latest session shows about 1 knt slower in ka72 than gps results and GPS Speed Reader. Also 5x10 is slower. Other categories match.
I have seen the same happening on last session posted by my team mate Julian on 14/11:
gpsteamchallenge.com.au/sailor_session/show?date=2021-11-14&team=33
I was wondering if this is happening to others, in case we could ask Dylan if there is a reason.
link: www.ka72.com/Track/t/482965
Also from GPSTeam challenge point of view, which one is the source of truth? I guess if the faster results are validated in other software I could edit it myself, but it might not be fair to the ones that only post with ka72
Yes I agree George, I have seen exactly the same thing (2sec and 5*10 slower in KA72). Perhaps the filtering has been changed and some runs are being ignored? The trouble with KA72 is that you cannot see which runs are being rejected from the results.
I generally use GPSSpeedreader these days, as it is quick, reliable and free.
I have just had a closer look at my track from 2-Dec on KA72.
The 2sec peak is shown as19.581, but the 100m speed is shown as 20.029, So something is not right.
The 5 ten second sections that are averaged to give the 5*10 appear to be selected from different parts of the track than those used in Speedreader.
I have notified Dylan.
Here is the section that ka72 picks for the top 2 sec:
ka72 picks a region 12 points after the highest speed for the top speed. Here is the correct region:
This does not look like a simple filter issue. There are a couple of points near the start with an acceleration above 2 m/s2, but the default threshold is 4.0, and dropping it to 2.0 still gives a result over 30 knots. The satellites and accuracy numbers for both regions are identical or nearly identical.
The top 10 second speed is also a bit further back in the track than it should be.
Clearly a problem with ka72.com. Having a higher speed for 100 m than for 2 seconds, as in tbwonder's example, is a clear indication that something is wrong with the ka72 calculations.
Both Julien and Dylan have been advised, it's strange that it only seems to happen with the latest mini firmware.
But all other analysis software aren't affected. As mentioned as far as I'm concerned you're better off using GPSSpeadreader and posting from that
It's free and runs on any platform, so nobody is disadvantaged.
It's also designed for the GPSTC in mind, the filters are set higher so eliminates more dodgy data. It also calculates the hour the way we think it should. you'll never get a 0hr with it, if you have a break in the hour, it will pick the best hour. In some cases this will be several knots higher than GPSResults and KA72
A number of sailors have mentioned this problem with KA72 when using the Motion. We cant find anything wrong with the motion data and Boardsurfr has shown.
The only conclusion that makes sense is that it is an issue with KA72. Others have also brought this to Dylans attention. I trust he is investigating it. 
All the tracks brought to my attention that differed in KA72 with either GPS-Speedreader or GPS-Results, gave the same, correct results in those other two programs.
In this case, Speed-reader and GPSresults are your "truth". 
I highly recommend downloading GPS-Speedreader, and checking your results in it if you are using a Motion.
thanks for the quick reply, will update my session and tell Julian to do the same for his.
The file came from the motion with screen, latest firmware.
The advantage of KA72 is that the public files are there for everyone to see, so gives an assurance of honesty to participants
Select to expand quoteboardsurfr said..>>>
Clearly a problem with ka72.com. Having a higher speed for 100 m than for 2 seconds, as in tbwonder's example, is a clear indication that something is wrong with the ka72 calculations.
Hmm so at 5hz 12 points is just over 2s.
I forget now, is that the difference between Earth and Sat time?
Is KA72 somehow mixing the two times?
The Mini Motion I was using has firmware version 3020
The latest on Motion website is 3037
Maybe it has to do with the change from sAcc tot sDOP in the latest firmware from the Motion (filter for sDOP in KA72).
Select to expand quotetbwonder said..
The Mini Motion I was using has firmware version 3020
The latest on Motion website is 3037
Ok so maybe I should say "recent" firmware.
As mentionned above, the logs I've been provided so far have shown expected and equal results on GPSSpeedReader/GPSAR/GPSResults.
Select to expand quoterp6conrad said..
Maybe it has to do with the change from sAcc tot sDOP in the latest firmware from the Motion (filter for sDOP in KA72).
It's hAcc to HDOP. I don't know how KA72 filters but the other platforms and Motion being in agreement tells me it's probably not an issue with the logs.Select to expand quotedecrepit said..
It appears only the latest firmware in the mini motions
No, someone raised this issue in August 2021 already.
Just looked at a few other of my recent tracks. The majority of them are the same in KA72 and GPSSpeedreader right down to the third decimal place in all six categories.
Select to expand quotedecrepit said..
Hmm so at 5hz 12 points is just over 2s.
I forget now, is that the difference between Earth and Sat time?
Is KA72 somehow mixing the two times?
We've seen problems with the second adjustments in the past, but that always has only affected the time that the speeds were reported at, not the actual speed. Which means it really only was an issue when comparing results from different programs, and looking at the times. But all these adjustments are always entire seconds. In this example, we had 2.4 seconds, so that can't be the problem. I also looked at the fastest 10 second run, and the offset between the ka72 and the correct result was a bit different.Select to expand quoterp6conrad said..
Maybe it has to do with the change from sAcc pDOP to sDOP hDOP in the latest firmware from the Motion (filter for sDOP in KA72).
Theoretically, that's possible, but the reported hDOP numbers for the region in question are too low to trigger a filter, and pretty much the same for the region that ka72 used and the region it should have used:
I think its fixed? KA72 and Speedreader agree exactly for me tonight. Mini Motion firmware 3034.
Select to expand quotejamesf said..
I think its fixed? KA72 and Speedreader agree exactly for me tonight. Mini Motion firmware 3034.
I don't think so James. Often the results are identical. But sometimes for a reason nobody yet has explained there is a difference.
I'm using a GW60 for on the water feedback, but uploading from minimotion through KA72 to GPSTC. So far I haven't seen anything unexpected.
I reported the potential problem to Dylan, Julien and the GPSTCin September. My experience has been as follows:
There has always been good correlation between the Mini Motion analysis in KA72 and GPSResults for the same session.
The issue appears to exist only in the LCD Motion. The analysis of LCD Motion logs on KA72 is often missing runs when compared to analysis by GPSResults and the on screen results of the unit. I only notice the error when the run is a good one.
There is always good correlation between the LCD Motion and Mini Motion when analysed through GPSResults. So I think the problem exists in how KA72 processes the LCD log files.
I'm currently testing the latest firmware version (3038) for Julien. I've sent to him the log files from both my Mini and LCD for two sessions over the weekend. He has indicated the log files are looking good. He has asked for more log files to confirm this initial positive review.
Interestingly, with this new firmware, there is good correlation in all disciplines between the analysis of logs files by both KA72 and GPSResults for both sessions so far. This could just be a coincidence. Analysis of more logs will tell.
Thanks John, I had a lower 10 sec average from lcd motion file on ka72 than gpsspeedreader yesterday, (3037), only looked as it's one of the gps results at our event yesterday, given the ka72 data is used for results others using motions might want to check theirs too
A bit confused to read this, the fact that this seems to be a known error since september and Julien is already working on a solution for this (that's good ...i think ??) is a bit dissapointing. We have seen previous "glitches" with Motions before and the only strategy around them seems to be to deny issues like hell , no transparency and after a software inprovement (why on earth????) present it like a non issue.....don"t get me wrong, there is no device on earth that's 100% but just be open about it .....
Select to expand quoteJohn340 said..
The issue appears to exist only in the LCD Motion. ....
The issue also happened to my teammate with a mini motion
Select to expand quoteXbraun54 said..
A bit confused ...
We found that the motion files and results analized by GPSResults and Speedreader are CORRECT, discrepancies are in KA72 processing.
Hope it is clear enough.
Select to expand quoteJohn340 said..
The issue appears to exist only in the LCD Motion.
No the issue I am seeing is from the Mini Motion.
As far as I can tell there is no problem with the Mini Motion files, they all look correct in Analysis software when compared to other tracks.Select to expand quoteXbraun54 said..
A bit confused to read this, the fact that this seems to be a known error since september and Julien is already working on a solution for this (that's good ...i think ??) is a bit dissapointing. We have seen previous "glitches" with Motions before and the only strategy around them seems to be to deny issues like hell , no transparency and after a software inprovement (why on earth????) present it like a non issue.....don"t get me wrong, there is no device on earth that's 100% but just be open about it .....
I needed a good laugh - thanks
Select to expand quoteXbraun54 said..
A bit confused to read this,
You certainly are, the problem seems to be more to do with KA72.
But there lies the problem, Julien says, the motion files work fine with all other software, and Dylan says all other hardware works fine in KA72. Where is the incentive for either of them to disrupt their busy schedule, get together and work out just what is happening.
"Julien and everybody else says, the motion files work fine with all other software."
Ignore XBraun54 / Jan Hendrik de Bruin from GP3S up here. He's moody for other reasons.
Select to expand quotedecrepit said..
Julien says, the motion files work fine with all other software, and Dylan says all other hardware works fine in KA72. Where is the incentive for either of them to disrupt their busy schedule, get together and work out just what is happening.
Well if there is no plan to get to the bottom of this quickly then I suggest that a note is put on GPSTC to suggest to people not to post using Motions with KA72. There are plenty of alternatives.
